We Don’t Know What’s Possible or What It Would Cost. Where Do We Start?

This is close to what one owner said to me after we had reviewed opportunities to deploy AI in his business. He simply could not connect any particular AI investment to a return he could defend to his partners.

That is the normal condition. It is a difficult question to answer. Owners are being asked to evaluate opportunities in a category where they have no comparable projects to draw on, no reference points from their own history, and mixed messages on where to begin.

There are three ways companies commonly start the process.

One option is to get your house in order if you have shadow AI. The term shadow AI means employees are using AI for company work outside anything the business has approved or can see. (Click here for additional information on shadow AI.) This can cause security concerns and significant problems with consistency or quality of work.

The second option for where to start is to get a strong foundation in place. That means holding off with those exciting AI pilot projects until you have audited your data and technology infrastructure to understand the practical limitations you face when it comes to adopting AI technology. It also addresses things like setting up a proper AI governance structure so your company can have proper oversight pertaining to AI use in the business. Change management is also a crucial step here because people make or break an AI implementation, so we need to take the temperature on the willingness and fears employees have around AI. We also need an audit of the high-value opportunities to deploy AI. Then with all that background, we can craft an appropriate AI strategy that integrates into your company’s current strategic direction. It’s a lot of work, but drastically improves the chances of success.

Some companies take an entirely different approach. They start with a single pilot project as a test. They want to experiment in a structured and meaningful way. If you go this route, it’s crucial to do this right. Are we sure we have the right test project in mind? Do we fully understand the implications on our business processes of deploying AI in this workflow? How will we measure if the effort is successful?

The right approach for your business depends on several factors, from appetite for change to the state of your technology infrastructure. Just be mindful as you take your first step. Sometimes, it pays to slow down for a while so you can go faster later.